摘要
针对粒子滤波存在粒子退化,会导致检测前跟踪(TBD)算法的检测和跟踪性能下降这一不足,提出了一种基于高斯-哈密顿滤波(GHF)高斯粒子滤波的TBD算法.该算法基于高斯粒子滤波,采用GHF算法构造的重要性密度函数采样连续出现粒子,考虑了最新的量测信息,采样粒子更逼近于真实的后验概率密度,克服了粒子退化问题.仿真结果表明:与基本TBD算法相比,所提出的TBD算法提高了对目标的检测和跟踪性能.
As the particle degeneracy problem in the particle filter,the detecting and tracking performance of track-before-detect（TBD）algorithm based on the particle filter descended.Therefore, the TBD algorithm based Gaussian-Hermite filter（GHF）Gaussian particle filter（GPF）was proposed.Based the GPF,the continue particles could be more approach to the real true posterior probability distribution including latest measuring information,which was sampled from the important density function based on the GHF.The experimental results show that the performance of detecting and tracking of targets by the proposed TBD algorithm is superior to the standard TBD algorithm.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2014年第3期23-27,32,共6页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(61179014
60872156)
关键词
检测前跟踪
粒子滤波
粒子退化
高斯粒子滤波
重要性密度函数
高斯-哈密顿滤波
track-before-detect（TBD）
particle filter
particle degeneracy
Gaussian particle filter（GPF）
important density function
Gaussian-Hermite filter（GHF
